Geomorphological and spectrophotometric analysis of Seth's circular niches on comet 67P/Churyumov-Gerasimenko using OSIRIS images

2020-10-26T16:06:45Z

Acceptance in OA@INAF

Geomorphological and spectrophotometric analysis of Seth's circular niches on

comet 67P/Churyumov-Gerasimenko using OSIRIS images

Title

LUCCHETTI, ALICE; PAJOLA, MAURIZIO; Fornasier, S.; Mottola, S.; Penasa, L.;

et al.

Authors

10.1093/mnras/stx1590

DOI

http://hdl.handle.net/20.500.12386/27994

Handle

MONTHLY NOTICES OF THE ROYAL ASTRONOMICAL SOCIETY

Journal

469

Number

Advance Access publication 2017 June 23

Geomorphological and spectrophotometric analysis of Seth’s circular

niches on comet 67P/Churyumov–Gerasimenko using OSIRIS images

A. Lucchetti,

M. Pajola,

S. Fornasier,

S. Mottola,

L. Penasa,

L. Jorda,

G. Cremonese,

C. Feller,

P. H. Hasselmann,

M. Massironi,

S. Ferrari,

G. Naletto,

N. Oklay,

H. Sierks,

C. Barbieri,

P. L. Lamy,

R. Rodrigo,

D. Koschny,

H. Rickman,

H. U. Keller,

J. Agarwal,

M. F. A’Hearn,

M. A. Barucci,

J. L. Bertaux,

I. Bertini,

S. Boudreault,

V. Da Deppo,

B. Davidsson,

S. Debei,

M. De Cecco,

J. Deller,

M. Fulle,

O. Groussin,

P. J. Gutierrez,

C. G¨uttler,

M. Hoffman,

S. F. Hviid,

W. H. Ip,

J. Knollenberg,

J. R. Kramm,

E. K¨uhrt,

M. K¨uppers,

L. M. Lara,

M. Lazzarin,

F. La Forgia,

L. Z. Lin,

J. J. Lopez Moreno,

F. Marzari,

F. Preusker,

F. Scholten,

X. Shi,

N. Thomas,

C. Tubiana

and J. B. Vincent

Affiliations are listed at the end of the paper

Accepted 2017 June 21. Received 2017 June 8; in original form 2017 March 29

A B S T R A C T

We provide a detailed geomorphological and spectrophotometric analysis of the circular niches located on the Seth region of 67P using OSIRIS images. The features can be related to landslide events that occurred on 67P and shaped its surface, as the recent Aswan cliff collapse detected in the same region. We therefore provide an analysis of the area pre- and post-perihelion suggesting that no specific changes have been observed. To assess this, after performing a geomorphological map of the area that allows us to identify different terrain units, we computed the boulders cumulative size frequency distribution (SFD) of the niches, before and after the perihelion passage. The niches SFDs are characterized by a similar trend with two different power-law indices within the same deposit: lower power-law value (between −2.3 and −2.7) for boulders smaller than 5 m and steeper power-law value (between −4.7 and −5.0) for boulders larger than 5 m. These trends suggest that smaller boulders have evolved and progressively have been depleted (lower power-law index), while bigger boulders are more representative of the event that generated the deposit and are less degraded. Then, we perform the spectrophotometric analysis of this region comparing pre- and post-perihelion results. We found colour changes within the area, in particular brighter patches related to the presence of exposed water ice mixed to the refractory materials have been detected in the post-perihelion images.

Key words: methods: data analysis – comets: general – comets: individual: 67P/Churyumov–

Gerasimenko.

1 I N T R O D U C T I O N

ESA’s Rosetta spacecraft orbited closely around comet 67P/Churyumov–Gerasimenko (hereafter 67P) during its 2 yr mis-sion. From 2014 August to 2016 September, the camera system

_{E-mail:alice.lucchetti@oapd.inaf.it}

Optical, Spectroscopic, and Infrared Remote Imaging System (OSIRIS; Keller et al.2007) onboard Rosetta observed the comet nucleus down to few centimetres per pixel providing detailed im-ages of the comet’s surface and activity. The study of 67P mor-phology has been performed with unprecedented detail thanks to the OSIRIS image spatial coverage and resolution. A first OSIRIS analysis on the nucleus structure, morphology and jets activity is re-ported in Sierks et al. (2015) and Thomas et al. (2015). The comet’s

Figure 1. Geological context of the Seth region and the circular niches analysed in this work. (a) NAC image acquired in 2014 August 6 showing regional

boundaries and nomenclatures of the comet 67P. The scale of the image is 2.2 m px−1 and the Seth region is coloured in red. (b) The same NAC image

displaying the circular niches analysed that are marked with the red square. (c) NAC image taken in 2014 August 21 (scale of 1.2 m px−1) showing a different

view of the circular niches, which are indicated by the red square. (d) The high-resolution WAC image acquired in 2016 August 24 (scale of 0.31 m px−1)

exhibiting the circular niches under study. We refer to the three niches as niches A, B and C as outlined in the image.

surface revealed a wide diversity in morphology such as layering (Massironi et al.2015), pits (Vincent et al.2015), boulders and frac-tures (Pajola et al.2015,2016c; El-Maarry et al.2015a), and high reflectivity boulders (Pommerol et al.2015). During the entire mis-sion, surface changes have been detected thanks to images acquired before and after perihelion passage (Groussin et al.2015; Pajola

2017; El-Maarry2017). The global spectrophotometric properties of the comet were investigated in detail by Fornasier et al. (2015) identifying three different groups of terrains depending on spectral slope values. In addition to a global characterization of the comet, OSIRIS allowed a detailed analysis of specific regions identified on the comet by means of high resolution and multifilter images (La Forgia et al.2015; Deshapriya et al.2016; Lucchetti et al.2016; Ok-lay et al.2016a; Pajola et al.2016d; Oklay et al.2017). We decided to follow the approach of characterizing at the highest resolution possible specific regions of interest (ROIs) on 67P. We therefore

performed a detailed analysis on an area belonging to the Seth re-gion (El-Maarry et al.2015b) that is characterized by flat-floored and steep-walled circular depressions (El-Maarry et al. 2015b; Giacomini et al.2016). We focused our attention on the circular niches of the Seth area (Fig.1) and performed a multidisciplinary study of these structures investigating their geomorphological and spectrophotometric properties. In addition, thanks to images ac-quired pre- and post-perihelion, we conducted a comparative anal-ysis to find if this area has been subjected to surface changes. This can be useful to provide constraints about Seth’s niches properties as well as their possible origin being one of the interesting features located on the surface of 67P. Specifically, the circular niches de-posits can be considered the result of landslide events that occurred on the comet surface, as the recent Aswan cliff collapse (Pajola

2017), where it has been reported the occurring of falling material from the adjacent cliff after the Rosetta perihelion passage. Hence,

with this work we plan to understand if the formation of these Seth’s circular niches occurred recently or if it is correlated to older events that have shaped the comet’s surface.

The paper is organized as follows. We perform a geomorphologi-cal analysis of the Seth circular niches identifying the different types of terrain units, then we derive the size frequency distribution (SFD) of boulders≥1.8 m located in the different considered niches. The overall analysis has been performed making use of the gravitational slope map and the erosion and insolation model of the area, which have been calculated on the high-resolution digital terrain model (DTM). In addition, we compare for one of the niches the boulders SFD derived in images acquired pre- and post-perihelion to observe if there have been any changes. Finally, we present a spectrophoto-metric analysis of the area before and after the Rosetta perihelion passage to describe the colour variation of the niches and to detect possible colour changes.

2 DATA S E T A N D M E T H O D S

The OSIRIS instrument was equipped with a narrow-angle camera (NAC), designed for the mapping of the comet’s surface with 12 filters covering the spectral range of 250–1000 nm in high spa-tial resolution, and a wide-angle camera (WAC), which was opti-mized for the gas and dust studies in the vicinity of the comet with a combination of 14 filters in the spectral range of 240–720 nm (Keller et al.2007). Our analysis of Seth’s niches was performed us-ing several OSIRIS images, includus-ing both NAC and WAC images, with different spatial resolutions. The WAC image used to obtain a geomorphological map of the area (Fig.1d) was taken in 2016 August 24, during one of the last elliptical orbits of Rosetta. This high-resolution image gives a unique view of the circular niches of the Seth region permitting the identification of different terrain units such as talus deposits, outcropping materials and gravitational accu-mulation deposits. It was taken at a distance of 3 km, resulting in a scale of 31 cm px−1and allowing the identification of features larger than 1 m located in the niches A and B (Fig.1d). In this image, there is also a smaller niche, niche C, whose orientation is not optimal for boulder counting; however, we consider it into the geomorpho-logical analysis and the spectrophotometric modelling in order to provide a complete description of the area. To determine if this area was affected by surface changes during the lifetime of Rosetta, we used a NAC image taken before perihelion on 2015 March 28, that allows us to count boulders in niche A and to compare the cumu-lative SFD with the one obtained after perihelion. We made use of

ARCGISsoftware to define the different terrain units characterizing

the three niches and to select the boulders, which are defined as positive relief detectable in various images obtained with different observation geometries, with the constant presence of an elongated shadow (if the phase angle is greater than 0◦; Pajola et al.2015). We selected boulders following the approach of Pajola et al. (2015,

2016b,c), i.e. by manually identifying these features as polygonal shapes and deriving their maximum length, i.e. the diameter and the corresponding area. The lowest boulders size considered for our analysis is a value above the three pixel sampling, to minimize the likelihood of misidentifications (Nyquist1928), even if we are able to identify also smaller features from their elongated shadow since the image was taken with a phase angle of 70◦. Anyway, we did not include boulders smaller than 5 px in the statistical analysis be-cause the slope of the SFD naturally drops off approaching the three pixel detection limit (Mazrouei, Daly & Barnouin2014; Michikami et al.2008). In addition, to obtain the cumulative SFD km−2, we

Figure 2. Gravitational slope map of Seth’s niches obtained using the SPG

(Preusker et al.2015) shape model and the gravitational potential model

derived assuming a homogeneous nucleus and accounting for the centrifugal force due to the nucleus rotation.

computed the area of each niche extracted from the 3D shape model of 67P (Preusker et al.2015).

To describe in detail the area, we computed the gravitational slope from the 3D shape model derived with the stereo-photogrammetry (SPG) technique (Preusker et al.2015) assuming a homogeneous nucleus with a density of 532 kg m−3(Jorda et al.2016), following the theory of Werner & Scheeres (1997) and accounting for the centrifugal force due to the comet rotation (Rossi & Fulchignoni

1999). The gravitational framework of the area (Fig.2) was obtained in order to correlate the slope with the identified terrain unit and possibly identify the location from where boulders fell down. The gravitational slope is defined as the angle between the local surface normal pointing inside the nucleus and the local acceleration vector, which comprises both gravitation and centrifugal accelerations. The accumulation deposits areas show slopes≤20◦, while outcropping walls show slopes≥50◦.

To support our analysis, a specific high-resolution DTM was cre-ated on the basis of the entire NAC data set that provided multiple stereo coverage of the cometary surface. The DTM was derived with the stereo-photoclinometry (SPC) technique (Jorda et al.2016) with a spatial resolution of 1 m (Fig.3). Moreover, on the basis of the high-resolution DTM, an insolation and erosion model was calcu-lated in order to have the total energy absorbed during a comet period and extrapolate the integrated erosion rate that characterizes niches A and B (Fig.4). The insolation model used for our analysis is described in Keller et al. (2015) and is derived approximating the shape of 67P’s circular niches through a polyhedron made of 300 000 triangular facets, which is obtained by decimating a high-resolution model based on images acquired during the orbital phase of Rosetta. The insolation of a facet is derived based on its ori-entation with respect to Sun. The erosion rate is referred to pure ice, however, the real erosion rate can be calculated considering the dust-to-ice ratio of the cometary material.

Finally, we consider a data set of images suitable to perform a detailed spectrophotometric analysis before and after the perihe-lion passage. The NAC images used are those acquired on 2014 November 11 at a distance of 44.8 km and on 2016 July 23 at a

Figure 3. High-resolution DTM derived for the circular niches under study.

The DTM was created with the SPC (Jorda et al.2016) technique with a

spatial resolution of 1 m.

Figure 4. Upper panel: erosion integrated over one orbit, the maximum pure-ice erosion is 0.53 m. Lower panel: insolation integrated over one

orbit, the maximum of total absorbed energy is 7.8 (J m−2).

distance of 7.8 km. The set of images used to derive spectrophoto-metric properties (that are in radiance units, Wm−2nm−1sr−1, at the central wavelength of each filter) are converted into radiance factor (I/F) by using I F = πIm Fr 2_, (1)

where Imis the measured intensity in radiance units, F_

π is the solar flux at 1 au measured at the central wavelength for each filter and r is the heliocentric distance of the comet in astronomical units. The images were reduced using the OSIRIS standard pipeline up to level 3B, following the data reduction steps (Tubiana et al.2015).

For the spectrophotometric analysis, the images of a given ob-serving sequence were first coregistered taking the NAC orange filter (centred at 649.2 nm) as reference and using a python script adapted from the code described in van der Walt et al. (2014) and the scikit-image library (van der Walt et al.2014). For post-perihelion images acquired at higher spatial resolution, we also used the opti-cal flow algorithm to improve the image coregistration (Farneback

2001). Then, images were photometrically corrected applying the Lommel–Seeliger disc law (Hapke1993) and the geometric infor-mation about the illumination and observation angles were derived using the SPC shape model (Jorda et al.2016). In the relative spec-trophotometry, the flux of a given region of interest is integrated over a 3×3 pixels box and normalized at 535 nm, following the same method described in Feller et al. (2016), Deshapriya et al. (2016), Fornasier et al. (2017) and Oklay et al. (2017).

The observation geometry, scale, phase angle and filters for each image used in the analysis are listed in Table1.

3 G E O M O R P H O L O G I C A L A N A LY S I S 3.1 Geomorphological map

The Seth region is located on the large lobe of the comet 67P and it is composed of a series of terraces alternated to steep walls show-ing outcrops of consolidated material (Giacomini et al.2016). The origin of semicircular terraces is still unknown and can be related to collapse of material tending to form talus and rockfall-like deposits (Giacomini et al.2016) or to the evolution of pits, the remnant of sinkhole collapse of buried cavities (Vincent et al.2015). Niches and pit rims are not clearly discernible in some cases. Indeed, some of Seth’s landforms can be the result of the interaction between the two origin processes: the original collapsed sinkhole occurred and the sublimation modified the landform shape with the conse-quent retreat of walls (Giacomini et al.2016). This is indicated by the gravitational accumulation deposits detected at the foot of the walls (Pajola et al. 2015; Vincent et al.2015). Between the dif-ferent circular features present in this area, we chose three niches that exhibit large gravitational accumulation deposits, probably due only to the collapse of material from the outcrops walls. A geomor-phological map of the area was performed highlighting the main geomorphological units, which are discriminated on the basis of the morphological properties following Giacomini et al. (2016) outlines (Fig.5).

On the steep wall of the outcropping terrain, several fractures have been detected. Steep walls are found in correspondence with high gravitational slopes (see Fig.2) larger than 40◦and usually show unconsolidated material at their foot. This is therefore inter-preted as the result of collapses due to erosion by sublimation. We can provide a first hint regarding the different units characterizing the area because of the strong correlation with the slope obtained by gravity, indeed, the gravitational slope map can be fundamental to understand the limits of the different terrains presented in the geomorphological map. For instance, we found that areas with a slope below 20◦are representative of boulders deposits, while the gravitational slope of the outcropping walls are steeper than 40◦, not allowing the boulders deposit to rest. In Fig.5, the different deposits can be identified as taluses and gravitational accumula-tions, separated on the basis of their texture. They have the same gravitational source and are interpreted as falling material caused by erosion of the adjacent cliffs. Talus deposits are generally con-stituted by finer material often showing some larger blocks located all along the foot of the walls (Giacomini et al.2016; Pajola et al.

Table 1. List of OSIRIS images used in this work. The last column reports the specific analysis performed on the corresponding images.

Date UT Distance (km) Scale (m px−1) Phase angle (◦) Filter ID λ(nm) Analysis

WAC 2016-08-24 20.58.01 3.1 0.31 70 F12 Red 629.8 Post-perihelion

geomorphological analysis and boulders counting

NAC 2015-03-28 6.49.49 29.3 0.55 68 F82 Orange 649.2 Pre-perihelion boulders counting

NAC 2014-11-14 22.07.23 44.8 0.84 90 F22 Orange 649.2 Pre-perihelion

spectrophotometric analysis NAC 2014-11-14 22.07.31 44.8 0.84 90 F23 Green 535.5 NAC 2014-11-14 22.07.39 44.8 0.84 90 F24 Blue 480.7 NAC 2014-11-14 22.07.47 44.8 0.84 90 F27 Hydra 701.2 NAC 2014-11-14 22.07.55 44.8 0.84 90 F28 Red 743.7 NAC 2014-11-14 22.08.03 44.8 0.84 90 F51 Ortho 805.3 NAC 2014-11-14 22.08.11 44.8 0.84 90 F61 Fe2O3 931.9

NAC 2016-07-23 07.30.48 7.8 0.15 89 F41 Near-IR 882.1 Post-perihelion

spectrophotometric analysis NAC 2016-07-23 07.31.00 7.8 0.15 89 F22 Orange 649.2 NAC 2016-07-23 07.31.11 7.8 0.15 89 F24 Blue 480.7 NAC 2016-07-23 07.35.36 7.8 0.15 89 F41 Near-IR 882.1 NAC 2016-07-23 07.35.48 7.8 0.15 89 F22 Orange 649.2 NAC 2016-07-23 07.35.59 7.8 0.15 89 F24 Blue 480.7

Figure 5. Geomorphological map of Seth’s niches presenting the identified units and lineaments. This map was derived by using the WAC image taken on 2016 August 24 at a distance of 3 km from the comet. The corresponding

scale is 0.31 m px−1.

2016d) and are thought to be the result of the erosion of the material fallen from the adjacent cliffs. The gravitational accumulation de-posits are composed by boulders of different dimensions and their source is well associated with the material detachment from the steep outcropping walls due to erosion by sublimation (La Forgia et al.2015; Massironi et al.2015; Oklay et al.2016a; Oklay et al.

2017).

3.2 Boulders analysis

Using the WAC image acquired on 2016 August 24, we counted boulders located in niches A and B. We do not consider niche C in this type of analysis since the orientation of the feature is not

optimal to perform boulders counting. The total number of boulders counted on niche A is 2526, 2216 of which have diameters larger than 1.8 m, while on niche B we counted 2313 boulders, with 2049 boulders larger than 1.8 m. We derived the cumulative SFD km−2using the corresponding area (0.08 and 0.06 km2_{for niches}

A and B, respectively) calculated from the 3D shape model of 67P (Preusker et al.2015). We then obtained the cumulative SFD of boulders using a constant bin size of 0.31 m, which is approximately the pixel resolution of the WAC image used for the analysis. We found different behaviour on the basis of the boulders dimension obtaining a similar trend for both niches. Specifically, for niche A, we found a power-law index of−2.3 +0.1/−0.1 for diameter ranging from 1.8 to 5 m (2047 boulders) and a power-law index of −5.0 +0.2/−0.1 (169 boulders) for boulders diameter larger than 5 m. The fitting regression line for niche B gives a power-law index of−2.7 +0.1/−0.1 for boulders diameter ranging from 1.8 to 5 m (1955 boulders) and a power-law index of −4.7 +0.2/−0.1 (94 boulders) for boulders diameter larger than 5 m. This behaviour is shown in Fig.6.

We outline that the fitting regression line used to obtain the power-law index does not consider those points that are equally cumula-tively repeated because they represent a poor statistic that has not been considered by the fit. This effect has been already observed for bigger boulder sizes as presented in Pajola et al. (2015) and Michikami et al. (2008). The lower limit of boulder dimension is assumed to be 5 px because the slope of the SFD naturally drops off approaching the three pixel detection limit (Mazrouei et al.2014; Michikami et al.2008). The cumulative number of boulders larger than 1.8 m is 27 000 km−2and 29 000 km−2for niches A and B, respectively.

We compare our results with what was found in previous works, in particular Pajola et al. (2015) defined three possible explanations for boulders formation on the basis of a first analysis of SFD ob-tained for the Northern hemisphere of 67P. In addition, there have been many works that tried to interpret the different boulders fields located in various regions or localized areas of the comet (La Forgia et al.2015; Pommerol et al.2015; Lucchetti et al.2016; Pajola et al.

2016a,c,d; Oklay et al.2016a,b; Oklay et al.2017). Generally, the behaviour of power-law indices reflects the process acting on the area analysed: (i) SFD with a power-law index ranging between

Figure 6. The upper panel shows the spatial distribution of boulders on niches A and B counted on the WAC image acquired on 2016 August 24. Boulders smaller than 5 px, where the boulder size corresponds to the length calculated from the corresponding area, are in green and are not taken into account in the statistic. The lower panel shows the cumulative SFD of boulders larger than 5 px identified on the two niches. The bin size is 0.31 m and vertical error bars

indicate the root of the cumulative number of counting boulders (as from Michikami et al.2008). The fitting regression line for niche A gives a power-law

index of−2.3 +0.1/−0.1 for boulders diameter ranging from 1.8 to 5 m and a power-law index of −5.0 +0.2/−0.1 for boulders diameter larger than 5 m. The

fitting regression line for niche B gives a power-law index of−2.7 +0.1/−0.1 for boulders diameter ranging from 1.8 to 5 m and a power-law index of −4.7

+0.2/−0.1 for boulders diameter larger than 5 m.

−5.0 and −6.0 is indicative of high fragmentation resulting from a depression formation with subsequent escape of high-pressure volatile, (ii) SFD with power-law indices between−3.5 and −4.5 have been observed to be indicative of gravitational events trig-gered by sublimation and/or thermal fracturing causing regressive erosion and (iii) SFD with power-law indices ranging between−1.0 and−2.0 are representative of the evolution of the original material formed during both the collapsing event or the gravitational event, not particularly renewed, or present in areas where continuous and high sublimation occurred or is still occurring. In some specific cases, different SFD power-law indices have been identified than those described above. This is interpreted as a possible mixture of different processes occurring on such areas (Deshapriya et al.2016). For circular niches, we expect a similar behaviour to the SFD of case (ii) because of the geomorphology of the area, however, we found a different behaviour for both niches. For smaller boulders (diameter ranging between 1.8 and 5 m) we found a power-law index of −2.3 and −2.7 for niches A and B, respectively. This value is between the two power-law indices of cases (ii) and (iii) suggesting that the boulders formed during the gravitational event

(usually characterized by a power-law index ranging from−3.5 to −4.5) have been evolved and progressively depleted lowering the power-law index of the SFD. In addition, the degradation of smaller boulders concurred to determine this behaviour. Hence, we can suggest that very degraded gravitational deposits show a lower power-law index maybe indicating that these deposits are older than the others. The power-law index describing the boulders larger than 5 m is instead characterized by a larger value,−5.0 and −4.9 for niches A and B, respectively, that can be related to the SFD case (i). The number of boulders considered in these SFDs is low, 169 and 94 boulders, for niches A and B, respectively, hence, the statistic can be affected by the small number of features counted. However, this different trend can be explained considering that bigger boulders are more representative of the event that raise the deposit and are less degraded than smaller boulders. Therefore, we expect higher power-law indices for bigger boulders and lower ones for smaller boulders within the same deposit, as found in our analysis.

We can also compare the results of niches A and B with what was found previously by Pajola et al. (2015) in the same area. The work reported the statistic analysis regarding boulders with a

Figure 7. (a) The NAC image acquired on 2015 March 28 (resolution scale of 0.55 m px−1) on which we perform the pre-perihelion analysis. The red box displays niche A on which we counted boulders. (b) The spatial distribution of boulders on niche A. Boulders smaller than 4 px, where the boulder size corresponds to the length calculated from the corresponding area, are in green and are not taken into account in the statistic. (c) The cumulative SFD of boulders larger than 4 px identified on niche A. The bin size is 0.55 m and vertical error bars indicate the root of the cumulative number of counting boulders (as from

Michikami et al.2008). The fitting regression line gives a power-law index of−2.7 +0.1/−0.1 for boulders diameter ranging from 2.7 to 5 m and a power-law

index of−4.9 +0.2/−0.1 for boulders diameter larger than 5 m.

diameter larger than 7 m that are characterized by a power-law in-dex of−6.5 and −5.5 for niches A and B, respectively. Pajola et al. (2015) explained the behaviour of these boulders field as the expres-sion of high fragmentation resulting from a depresexpres-sion formation with subsequent escape of high pressure volatile [SFD case (iii)]. The existing difference with our work (−5.0 and −4.9 for boulders larger than 5 m for niches A and B) can be explained by the fact that increasing the resolution of the image, the boulders counting improves allowing the identification of smaller boulders that en-riched the statistic and slightly changed the SFD trend. To assess if there have been any changes in this area, we considered a NAC image taken before perihelion to have a direct comparison between the boulders identified. For this purpose, we took an image acquired on 2015 March 28 (see Table1) that has a similar geometry of the WAC image, but contains only niche A. Therefore, we performed the boulder analysis identifying 1329 boulders (Fig.7). This number is smaller than the previous one because of the different resolution. Considering a bin size of 0.55 m (equal to the image pixel resolution as in the previous case), the fitting regression line has a power-law

index equal to−2.7 +0.1/−0.1 for boulders diameter ranging from 2.7 to 5 m (974 boulders) and a power-law index of−4.9 +0.2/−0.1 for boulders diameter larger than 5 m (225 boulders). Below 4 px the distribution starts to roll over indicating that the boulder distribu-tion is not complete. The number of boulders with a diameter larger than 2.7 m is 1199 and it is comparable with the number of boul-ders larger than 2.7 m in niche A of the WAC image, equal to 1148. The SFD power-law indices obtained in images acquired pre- and post-perihelion are slightly in agreement, even if the SFD power-law index obtained with WAC image is little lower. This implies no particular changes in the area (in particular in niche A), at least from the geomorphological point of view and, in addition, it suggests that the source event of these gravitational deposits occurred in the past and is not related to the 67P recent detected activity, as found for Aswan cliff (Pajola2017). This consideration is also supported by the results coming from the erosion and insolation model integrated over one 67P orbit (Fig.4): the erosion rate characterizing the area has a maximum of only 0.53 m and does not strongly affect the area. However, it must be noted that the erosion mainly affects one

Figure 8. (a) The image acquired on 2014 November 11 used to perform the pre-perihelion analysis. The image was taken at a distance of 44.8 km, resulting

in a scale of 0.84 m px−1. The area containing Seth’s niches is outlined by the red box and its close out is shown in (b). (c) Spectral slope of the niches area

computed in the 535–931 nm range, after the normalization at 535 nm and it is in per cent (100 nm)−1unit. Sun is towards the top and the phase angle is 90◦.

specific side of niches A and B from which we expect the material fell down in the past. In the last 67P orbit no gravitational collapses have been observed, but the constant presence of this selective ero-sion can result in future collapses coming from that specific side of the niches.

4 S P E C T R O P H OT O M E T R I C A N A LY S I S

We performed a spectrophotometric analysis on NAC images ac-quired before and after perihelion in order to characterize the colour changes after the Rosetta perihelion passage. For the pre-perihelion study, we chose a colour sequence centred on the Seth niches that was acquired on 2014 November 14 at phase angle of 90◦, at a he-liocentric distance of 2.97 au and with seven NAC filters (F22, F23, F24, F27, F28, F51 and F61). The distance between Rosetta and the comet at that time was 44.8 km, giving a spatial resolution of 0.84 m px−1with the NAC camera. We evaluated the spectral slope in the 535–931 nm range normalizing at 535 nm (Fig.8). The colour bar was chosen so that blue colour corresponds to regions with a lower slope than average, while the red colour represents regions with a higher slope than average. The average value obtained from

a Gaussian fitting of the slope distribution is 16.6± 1.4 [ per cent (100 nm)−1]. This value can be compared with the spectral slope analysis presented in Fornasier et al. (2015,2016), even if it must be noted that in these studies the spectral slope was evaluated in the 535–882 nm range, but in the data set of 2014 November the 882 nm filter observation was missing. However, considering that the spec-trum is usually linear in the 882–931 nm range and that the slope is evaluated normalizing at the same wavelength (535 nm), the varia-tions related to the slightly different wavelength range used should be negligible. Fornasier et al. (2015) analysed the spectral slope from colour images acquired during 2014 July–August for the phase angle lower than 55◦. They reported an important phase reddening for the nucleus of comet 67P, as well as colour heterogeneities at macro scale with the identification of three kinds of terrains: (i) terrains with the lower spectral slope [11–14 per cent (100 nm)−1 at phase angle 54◦) mostly associated with regions showing ac-tivity and having a surface composition enriched in water ice, (ii) average slope terrains [having a spectral slope of 14–18 per cent (100 nm)−1at phase angle 54◦] and (iii) redder terrains [slope>18 per cent (100 nm)−1at phase angle 54◦] associated mostly to regions covered by dust. The Seth region was identified as belonging to the

Figure 9. On the left are shown some selected ROIs of the pre-perihelion where the spectrophotometry has been computed. On the right, the relative reflectance for the corresponding selected regions is displayed, after normalization at 535 nm. The black filled circle on the right is represented by a white not filled circle on the image on the left.

lower spectral slope group in Fornasier et al. (2015) analysis. More detailed analysis of Seth in high spatial resolutions are reported in Oklay et al. (2016b), where it is shown that this region is mainly active even if a large area is possibly mixed with the airfall material described in Thomas et al. (2015). The fact that Seth is relatively bluer is confirmed also by comparing the spectral slope here derived atα = 90◦[16.6 per cent (100 nm)−1] with the one extrapolated at the same phase angle from the spectral reddening coefficients derived from 2014 July–August observations of the comet in the phase angle range 0◦–55◦[>19 per cent (100 nm)−1, see fig. S3 in Fornasier et al.2016.

The comparison of the spectral slope map with the geomorpho-logical units reveals a strong correlation with the morphogeomorpho-logical units identified, similarly to what was found in La Forgia et al. (2015). Indeed, Fig.8shows some local heterogeneities in the spec-tral slope value, with rough terrains having gravitational accumu-lation deposits which are spectrally redder, while talus deposits at the feet of the terraces are spectrally bluer. We thus investigate 16 different ROIs, as shown in Fig.9. The ROIs represented by the blue asterisk/cyan triangle symbols in Fig.9a and by the blue circle and cyan triangle in Fig.9b are clear examples of relatively bluer terrains, associated with taluses deposits.

The spectrophotometry indicates a linear increasing of the re-flectance with the wavelength, without any evidence of clear ab-sorption bands. However, an enhancement of the reflectance at 700 and 743 nm is evident, especially in the shadowed area (as for the ROI represented by the blue circle in Fig.9a, or red star in Fig.9b). A similar spectral behaviour was observed by Fornasier et al. (2015) and attributed to cometary emissions (faint but visible especially on shadows on the nucleus) from the coma between the spacecraft and the 67P’s nucleus. This implies that even at large heliocentric dis-tance the comet was already active. The spectral slope values in the 535–931 nm range for the different ROIs investigated are shown in Table2.

The Seth region was also observed after the perihelion passage at a similar phase angle (89◦) and a similar large outbound heliocentric distance (3.44 au) on 2016 July 23. At this time the spacecraft was much closer to the comet, at an altitude of 7.8 km giving a spatial resolution of about 0.15 m. During these close observations, a couple of colour sequences were acquired at UT time 7h31 and 7h35 using three NAC filters centred at 480, 649 and 882 nm. Fig.10shows the RGB maps in false colours of the Seth region and the associated spectral slope maps. As most of the previous results on spectrophotometry of the 67P comet are normalized at

Table 2. Spectral slope values for the ROIs investigated in Figs9(535–

931 nm range),11and12(535–882 nm range).

Figure ROIs Spectral slope [ per cent (100 nm)−1]

9a White circle 16.9± 0.3 9a Red star 17.1± 0.3 9a Blue asterisk 15.8± 0.3 9a Green triangle 17.2± 0.3 9a Magenta square 16.1± 0.3 9a Orange triangle 18.3± 0.3

9a Cyan upsidedown triangle 14.8± 0.3

9a Royal blue circle 17.9± 0.3

9b White circle 18.4± 0.3 9b Red star 17.0± 0.3 9b Blue asterisk 17.8± 0.3 9b Green triangle 18.0± 0.3 9b Magenta square 15.5± 0.3 9b Orange triangle 16.6± 0.3

9b Cyan upsidedown triangle 15.4± 0.3

9b Royal blue circle 14.7± 0.3

11a White circle 16.2± 0.4 11a Red star − 0.8 ± 0.4 11a Blue asterisk 18.0± 0.4 11a Green triangle 16.9± 0.4 11a Magenta square 10.9± 0.4 11b White circle 14.8± 0.4 11b Red star 7.6± 0.4 11b Blue asterisk 9.2± 0.7 11b Green triangle 14.9± 0.4 11b Magenta square 0.9± 0.7 12a White circle 15.5± 0.4 12a Red star 13.1± 0.4 12a Blue asterisk 3.2± 0.4 12a Green triangle 14.6± 0.4 12a Magenta square 15.7± 0.4 12b White circle 16.1± 0.4 12b Red star 15.5± 0.4 12b Blue asterisk 17.3± 0.7 12b Green triangle 9.1± 0.4 12b Magenta square 6.0± 0.7

535 nm, we evaluate the spectral slope in the 882–535 nm range normalizing our data at 535 nm using a linear interpolation between the 480 and 649 nm observations. Compared to the pre-perihelion images, several bright and bluer regions close to shadows or in the shadows are evident, indicating a local enrichment in ice mixed to the refractory material.

We thus investigate the spectrophotometry of some ROIs (Figs11and12) normalizing our data at 535 nm, as done for the spectral slope maps. Several points clearly show a flat spectrophoto-metric behaviour or a spectrophotometry with spectral slope values much lower than the surrounding regions. In particular, the ROI in-dicated by the red star symbol in Fig.11a, by the magenta square in Fig.11b and by the blue asterisk in Fig.12a are spectrally bluer and they are consistent with the presence of exposed water ice mixed to the refractory materials, as previously observed in different re-gions of the nucleus (Fornasier et al.2015; Pommerol et al.2015; Barucci et al.2016; Deshapriya et al.2016; Filacchione et al.2016; Fornasier et al.2016; Oklay et al.2016a,b; Pajola2017; Oklay et al.

2017). The spectral slope values in the 535–882 nm range for the different ROIs investigated are shown in Table2.

Despite the higher abundance of bluer regions close to shadows compared to the pre-perihelion images, the average spectral slope has not changed significantly even if it has a higher dispersion. We find a slope of 16.8± 2.1 [per cent (100 nm)−1] and 16.6± 2.6 [per cent (100 nm)−1] for the 7h31 and 7h35 images, respectively. These results are in agreement with the ones found by Fornasier et al. (2016), who reported a progressive blueing of the 67P nu-cleus’ colours approaching the perihelion compared to the inbound observations at large heliocentric distance, followed by a reddening at post-perihelion distances>2 au. They interpreted this contin-uous colour changes of the nucleus along its orbit as due to the progressive thinning of the dust mantle when the comet approached perihelion and became more active, partially exposing the underly-ing ice-enriched layers. In the post-perihelion phase, the cometary activity did not sustain removal of dust, and the thicker dust mantle covered and masked the spectrally bluer ice-enriched layers, giving a nucleus spectrophotometry with spectral slope as red as the one observed in the pre-perihelion phase at high heliocentric distance. A similar trend was recently observed also in the Southern hemisphere region Anhur, with a steep increase of the spectral slope in pre- and post-perihelion phase at large heliocentric distances compared to the observations close to the perihelion passage (Fornasier et al.

2017).

5 C O N C L U S I O N S

This work focused on the circular niches located in the Seth region on the body of 67P. By means of OSIRIS NAC and WAC images, we have performed a detailed analysis of the area using different resolutions. During one of the latest Rosetta elliptical orbits, a set of images were acquired giving a unique view of these interesting features at high resolution. Specifically, thanks to the WAC image taken on 2016 August 24 with a scale of 0.31 m px−1, we made a geomorphological map of the area identifying different terrain units. These niches are characterized by the presence of both gravitational accumulation deposits and talus deposits that can be separated on the basis of their texture. Such deposits cover the terrain outlined by the adjacent outcropping walls, which are defined as another geomorphological unit.

Between the different features that can be identified in the image, the boulders field deposits mainly characterize the interior of the circular niches. Thanks to the scale of the images we were able to identify boulders with different shapes with a diameter larger than 1 m. After the identification of these features, we statistically analysed their distribution deriving a cumulative SFD with similar power-law indices for niches A and B. The fitting regression lines returned two different power-law indices on the basis of the boulders dimension within the same deposit. Indeed, niche A is described by a power-law index equal to−2.3 for boulders diameter ranging from 1.8 and 5 m, and a power-law index of−5.0 for boulders larger than 5 m. Niche B is described by a similar behaviour consisting of a power-law index of−2.7 for boulders diameter ranging from 1.8 and 5 m, and a power-law index of−4.7 for boulders larger than 5 m. When comparing these results with the Pajola et al. (2015) work, we obtain that the values for smaller boulders (smaller than 5 m) fall within the SFD case (ii) and (iii), while larger boulders (larger than 5 m) can be related to the SFD case (i). This suggests that smaller boulders have formed during gravitational collapses, but have been progressively degraded and depleted lowering the power-law index and, in addition, the degradation of smaller boulders might have concurred to determine this SFD trend. The behaviour found for larger boulders, that are located in proximity of highly fractured

Figure 10. Upper panel: RGB images, in false colours, of the investigated region for images taken on 2016 July 23 with filters centred on 480, 649 and

882 nm. Rosetta was at a distance of 7.9 km resulting in a scale of 0.15 m px−1. On the left is the NAC image acquired at UT time 07h31 showing parts of

niches A, B and C, while on the left is the NAC image acquired at UT time 07h35 displaying the niches A and C. Some frosts/water ice enriched regions are seen close to the shadows or in shadowed regions (regions with relatively bluer colours). Lower panel: spectral slope of the corresponding images at a phase

angle of 89◦. The slope is computed in the 535–882 nm range, after the normalization at 535 nm and it is in per cent (100 nm)−1.

walls, can be the expression of an enhanced fragmentation, which might be produced during the depression formation that caused the release of pressurized volatiles. To detect if there have been any geomorphological changes, we used a NAC image taken before perihelion in 2015 March to compare the distribution of boulders within the niches. In particular, we found similar pre/post results for niche A implying no specific changes in the area. In addition, from the erosion model characterizing the niches it must be noted that

the area is not affected by strong erosion over one comet period, hence, we did not expect large variations on the different geological features. After identifying multiple terrains with different kinds of textures on Seth’s niches, we performed the spectral analysis of this region in order to understand if there are any spectrophotometric differences between the various geomorphological units. We per-formed the spectrophotometric analysis for pre- and post-perihelion images to detect possible colour changes in the area. For this

Figure 11. On the left are shown 10 ROIs of the NAC 07h31 UT where the spectrophotometry has been computed. On the right, the relative reflectance for the corresponding selected regions is displayed. As the green filter is not available, the normalization is done by linear fitting and interpolation using the 649 and 480 nm filter images. The black filled circle on the right is represented by a white circle on the image on the left.

purpose, we evaluated the spectral slope for data sets acquired in 2014 November and 2016 July, whose images have different resolu-tions (0.84 and 0.15 m for the pre- and post-perihelion, respectively) but similar geometry. We also identified and analysed different ROIs that sample many areas of the niches (16 ROIs for pre-perihelion data set and 20 ROIs for post-perihelion data sets). Comparing the results, we found that the average spectral slope has not changed significantly after the Rosetta perihelion passage, but we observed bluer spots in images taken after perihelion indicating the presence of exposed water ice mixed to the refractory materials, in agreement with what was previously observed in different regions of the nu-cleus (Fornasier et al.2016; Oklay et al.2016a; Oklay et al.2017; Pajola2017).

The overall analysis implies minor colour changes in images with a spatial resolution of 15 cm px−1, while geomorphological variations have not been observed, maybe because the expected erosion was below the detection limit, i.e. the spatial resolution of images used for geomorphological analysis. No specific changes in boulders deposits suggest that the gravitational event that gave birth to these deposits is not related to the recent detected activity of 67P. This means that the landslide that originated the boulders deposits

occurred in the past, maybe implying that these areas can be older than the nearby ones. This is also in agreement with El-Maarry (2017)’s work in which the analysis of detected surface changes imply a more active comet in the past.

AC K N OW L E D G E M E N T S

Optical, Spectroscopic and Infrared Remote Imaging System (OSIRIS) was built by a consortium of the Max-Planck-Institut f¨ur Sonnensystemforschung, in G¨ottingen, Germany, CISAS-University of Padova, Italy, the Laboratoire d’Astrophysique de Marseille, France, the Instituto de Astrofisica de Andalucia, CSIC, Granada, Spain, the Research and Scientific Support Department of the European Space Agency, Noordwijk, the Netherlands, the In-stituto Nacional de Tecnica Aeroespacial, Madrid, Spain, the Uni-versidad Politechnica de Madrid, Spain, the Department of Physics and Astronomy of Uppsala University, Sweden and the Institut f¨ur Datentechnik und Kommunikationsnetze der Technischen Univer-sit¨at Braunschweig, Germany. The support of the national funding agencies of Germany (DLR), Italy (ASI), France (CNES), Spain (MEC), Sweden (SNSB) and the ESA Technical Directorate is

Figure 12. On the left are shown 10 selected ROIs of the NAC 07h35 UT where the spectrophotometry has been computed. On the right, the relative reflectance for the corresponding selected regions is displayed. As the green filter is not available, the normalization is done by linear fitting and interpolation using the 649 and 480 nm filter images. The black filled circle on the right is represented by a white circle on the image on the left.

gratefully acknowledged. We thank the ESA teams at ESAC, ESOC and ESTEC for their work in support of the Rosetta mission. We made use of theARCGIS10.2 software together withIDLandMATLAB

software to perform the presented analysis.

R E F E R E N C E S

Barucci M. A. et al., 2016, A&A, 595, A102 Deshapriya J. D. P. et al., 2016, MNRAS, 462, S274 El-Maarry M. R. et al., 2017, Science, 355, 1392

El-Maarry M. R. et al., 2015a, Geophys. Res. Lett., 42, 5170 El-Maarry M. R. et al., 2015b, A&A, 583, A26

Farneback G., 2001, Proc. Eighth IEEE Int. Conf. on Comput. Vis. ICCV 2001, Very High Accuracy Velocity Estimation Using Orientation Ten-sors, Parametric Motion, and Simultaneous Segmentation of the Motion Field. IEEE Computer Society Press, Vancouver, BC, p. 171

Feller C. et al., 2016, MNRAS, 462, S287 Filacchione G. et al., 2016, Nature, 529, 368 Fornasier S. et al., 2015, A&A, 583, A30 Fornasier S. et al., 2016, Science, 354, 1566 Fornasier S. et al., 2017, MNRAS, 469, S93 Giacomini L. et al., 2016, MNRAS, 462, S352

Groussin O. et al., 2015, A&A, 583, A36

Hapke B., 1993, Theory of Reflectance and Emittance Spectroscopy. Cambridge Univ. Press, Cambridge

Jorda L. et al., 2016, Icarus, 277, 257

Keller H. U. et al., 2007, Space Sci. Rev., 128, 433 Keller H. U. et al., 2015, A&A, 583, A34 La Forgia F. et al., 2015, A&A, 583, A41 Lucchetti A. et al., 2016, A&A, 585, L1 Massironi M. et al., 2015, Nature, 526, 402

Mazrouei S., Daly M. G., Barnouin O. S., Ernst C. M., DeSouza I., 2014, Icarus, 229, 181

Michikami T. et al., 2008, Earth Planets Space, 60, 13 Nyquist H., 1928, 47, 617

Oklay N. et al., 2016a, MNRAS, 462, S394 Oklay N. et al., 2016b, A&A, 586, A80 Oklay N. et al., 2017, MNRAS, 469, S582 Pajola M. et al., 2017, Nature Astron., 1, 0092 Pajola M. et al., 2015, A&A, 583, A37 Pajola M. et al., 2016a, MNRAS, 462, S242 Pajola M. et al., 2016b, A&A, 585, A85 Pajola M. et al., 2016c, A&A, 592, L2 Pajola M. et al., 2016d, A&A, 592, A69 Pommerol A. et al., 2015, A&A, 583, A25

Preusker F. et al., 2015, A&A, 583, A33

Rossi A., Fulchignoni M., 1999, Planet. Space Sci., 47, 873 Sierks H. et al., 2015, Science, 347, aaa1044

Thomas N. et al., 2015, Science, 347, aaa0440 Tubiana C. et al., 2015, A&A, 583, A46

van der Walt S. et al., 2014, preprint (arXiv:1407.6245)

Vincent J.-B. et al., 2015, Nature, 523, 63

Werner R. A., Scheeres D. J., 1997, Celest. Mech. Dyn. Astron., 65, 313

1_{INAF - OAPD Astronomical Observatory of Padova, Vic. Osservatorio 5,} I-35122 Padova, Italy

2_{NASA Ames Research Center, Moffett Field, CA 94035, USA}

3_{LESIA-Observatoire de Paris, PSL Research University, CNRS, Univ.} Paris-Diderot, Sorbonne Paris Cite, UMPC Univ. Paris 06, Sorbonne uni-versites, 5 Place J. Janssen, F-92195 Meudon Principal Cedex, France 4_{Deutsches Zentrum f¨ur Luft- und Raumfahrt (DLR), Institut f¨ur} Planeten-forschung, Rutherfordstrasse 2, D-12489 Berlin, Germany

5_{Center of Studies and Activities for Space, CISAS, ‘G. Colombo’, University} of Padova, Via Venezia 15, I-35131 Padova, Italy

6_{Laboratoire d’Astrophysique de Marseille, UMR 7326 CNRS &} Aix-Marseille Universit´e, 38 rue Fr´ed´eric Joliot-Curie, F-13388 Aix-Marseille cedex 13, France

7_{Geosciences Department, University of Padova, Via G. Gradenigo 6,} I-35131 Padova, Italy

8_{Department of Information Engineering, University of Padova, Via} Gradenigo 6/B, I-35131 Padova, Italy

9_{CNR-IFN UOS Padova LUXOR, Via Trasea 7, I-35131 Padova, Italy} 10_{Max-Planck-Institut f¨ur Sonnensystemforschung, Justus-von-Liebig-Weg,} 3 D-37077 G¨ottingen, Germany

11_{Department of Physics and Astronomy ‘G. Galilei’, University of Padova,} Vic. Osservatorio 3, I-35122 Padova, Italy

12_{Centro de Astrobiologa, CSIC-INTA, E-28850 Torrejon de Ardoz, Madrid,} Spain

13_{International Space Science Institute, Hallerstrasse 6, CH-3012 Bern,} Switzerland

14_{Scientific Support Office, European Space Research and Technology} Cen-tre/ESA, Keplerlaan 1, Postbus 299, NL-2201 AZ Noordwijk ZH, the Nether-lands

15_{Department of Physics and Astronomy, Uppsala University, E-75120} Up-psala, Sweden

16_{PAS Space Research Center, Bartycka 18A, PL-00716 Warszawa, Poland} 17_{Institute for Geophysics and Extraterrestrial Physics, TU Braunschweig,} D-38106 Braunschweig, Germany

18_{Department for Astronomy, University of Maryland, College Park,} MD 20742-2421, USA

19_{LATMOS, CNRS/UVSQ/IPSL, 11 Boulevard d’Alembert, F-78280} Guyan-court, France

20_{NASA Jet Propulsion Laboratory, 4800 Oak Grove Drive, Pasadena, CA} 91109, USA

21_{Department of Mechanical Engineering, University of Padova, Via Venezia} 1, I-35131 Padova, Italy

22_{UNITN, University of Trento, Via Mesiano, 77, I-38100 Trento, Italy} 23_{INAF Osservatorio Astronomico di Trieste, Via Tiepolo 11, I-34143} Tri-este, Italy

24_{Instituto de Astrofisica de Andalucia CSIC, Glorieta de la Astronomia,} E-18008 Granada, Spain

25_{Institute for Space Science, National Central University, 32054 Chung-Li,} Taiwan

26_{Operations Department European Space Astronomy Centre/ESA, PO Box} 78, E-28691 Villanueva de la Canada, Madrid, Spain

27_{Physikalisches Institut der Universit¨at Bern, Sidlerstr. 5, CH-3012 Bern,} Switzerland